Horning, SJ

8038 Background: P21 is a downstream effector protein of p53 that induces cell cycle arrest by inhibiting cyclin/cyclin dependent kinase (cdK) complexes. In this prospective correlative study, we investigated the prognostic significance of p21, p53, Bcl-6, and Bcl-2 in diffuse large B cell lymphoma (DLBCL) in the context of E4494, a randomized trial comparing conventional CHOP to rituximab(R)-CHOP induction, and maintenance R (MR) to observation for responding patients. Methods: Protein expression was quantified by immunohistochemical (IHC) staining of 198 DLBCL paraffin-embedded biopsy specimens and scored by an expert panel of hematopathologists. Results: No differences in the distribution of patient characteristics were detected between p21+ and p21- cases and between p53+ and p53- cases as well as across treatment arms. Among all study patients, there were no differences in clinical outcomes between p21+ and p21- cases. However, when analyzed by induction arm (removing the effect of MR), R-CHOP patients had marginally better FFS (p=0.06) if p21+; among CHOP-treated patients, p21 status had no effect on outcome. For R-CHOP patients but NOT CHOP patients, p21+ was an independent, favorable prognostic factor after adjusting for Bcl-6, IPI and Bcl-2 in multivariate analysis (FFS relative risk 0.3; p=0.003; OS relative risk 0.4; p=0.02). P21+ patients treated with R-CHOP had higher %FFS at 5 years compared to p21+ patients treated with CHOP (61 ± 7% vs. 24 ± 7%; p=0.007) but p21- R-CHOP and CHOP patients had similar FFS (37 ± 7% vs. 35 ± 7%; p=0.64). P53 staining (+ scored as either >20% or 50%) did not predict for FFS or OS in uni- or multivariate analyses. The p53+/p21- phenotype, a possible surrogate for mutated p53, was not prognostic. Further, no significant correlation was seen between p53 and p21 expression by IHC. Conclusions: These data suggest that p21 expression by IHC predicts for favorable outcome in older DLBCL patients treated with R-CHOP but not in those treated with CHOP. Furthermore, p21+ identifies patients who benefit from the addition of rituximab to CHOP. Complex interactions between p53, p21, and Bcl-6, as well as other unknown pathways, may account for this treatment-related effect and should be further investigated. [Table: see text]

8004 Background: To determine optimal induction and the role of maintenance, the E1496 study randomized patients (pts) to induction CVP (cyclophosphamide, vincristine, prednisone) versus CF (cyclophosphamide 1 G/m2 d1, fludarabine 20 mg/m2 d1–5 every 28 d) for 2 cycles beyond best response (maximum 8). Responding and stable pts were secondarily randomized to MR (375 mg/m2 weekly × 4 every 6 months for 2 years [yr]) or observation (OBS). Methods: Due to early deaths the CF arm was closed to accrual with 115 pts randomized to CF and 119 pts to CVP (thereafter all pts were assigned to CVP prior to maintenance randomization). The results presented here compare the outcome of CF patients with the subset of E1496 pts randomized to CVP (CVPR). Results: Median follow-up on pts randomized to induction is 6.5 yr. Toxic deaths occurred in 8 (7%) CF pts during induction and 4 additional deaths (1 OBS, 3 MR) occurred among the 69 (6%) CF pts randomized to MR or OBS. Causes of death were infection (9), liver failure (2), CNS gliosis (1). CF pts received a median of 5 cycles compared to 7 cycles for CVPR. The CR rate was 51% vs 22% (p=0.00001) and the PR rate was 35% vs 55% for CF vs. CVPR, respectively. Four-yr PFS for CF vs. CVPR was 49% vs 45% (p=0.19) and OS was 66% vs. 81% (p=0.12), respectively. Of 45 CF deaths, 23 (51%) occurred without lymphoma progression compared to 5 (13%) of 38 CVPR deaths (p=0.0004). More than 90% of CF patients randomized to maintenance achieved protocol-defined minimal residual disease compared with 64% CVPR pts. Maintenance therapy had no impact on 2 yr PFS for the 67 evaluable randomized CF pts, which was 74% for MR vs. 73% for OBS (p=0.19). In contrast, 2 yr PFS was 73% for MR and 42% for OBS in randomized CVPR pts (p=0.004). Survival at 2 yr for MR vs OBS was: CF 79% vs 91% (p=0.19) compared with CVP 98% vs 93% (p=0.21). Conclusions: Induction with CF results in higher CR and miminal residual disease rates than CVP. However, gains in remission quality with CF (in the dose and schedule used here) were offset by early and late deaths in the absence of progressive lymphoma. In E1496, the benefit of MR was influenced by the induction chemotherapy. [Table: see text]